Taming Disorder in Pharmaceutical Solids

The emergence of non-crystalline materials during pharmaceutical development is potentially a major problem due to their typically very different physical properties. Although these properties, especially solubility, can be favourable, the difficulty of characterizing disordered materials means that they are generally avoided.

The Durham NMR group have helped pioneer the developing field of "NMR crystallography" which combines experimental solid-state NMR spectroscopy with computational calculations. The calculations have traditionally been limited to crystalline materials, but new "fragment"-based approaches that do not require periodicity should allow us to predict NMR properties for disordered materials. We can use new machine learning methods to build models that would allow us to predict NMR spectra for complex, real-life materials.

The project will also develop new experimental methodologies to characterise disordered materials. The broad features of conventional NMR spectra of non-crystalline materials limits our ability to distinguish different disordered materials. A number of NMR techniques can (quite literally) add another dimension of discrimination, and other exciting developments include "Dynamic Nuclear Polarisation" (DNP), which can boost NMR signals by orders of magnitude and open up two-dimensional NMR techniques that are normally impossible. This experimental data, combined with the computational work should allow to identify the conformations of molecules in amorphous materials, something which is currently impossible. As we have shown with other drug systems, understanding the molecular origin of disorder can explain why some disordered materials are safe to progress in pharmaceutical development.